Machine learning enables detection of early-stage colorectal cancer by whole-genome sequencing of plasma cell-free DNA.

BACKGROUND: Blood-based methods using cell-free DNA (cfDNA) are under development as an alternative to existing screening tests. However, early-stage detection of cancer using tumor-derived cfDNA has proven challenging because of the small proportion of cfDNA derived from tumor tissue in early-stage disease. A machine learning approach to discover signatures in cfDNA, potentially reflective of both tumor and non-tumor contributions, may represent a promising direction for the early detection of cancer. METHODS: Whole-genome sequencing was performed on cfDNA extracted from plasma samples (N = 546 colorectal cancer and 271 non-cancer controls). Reads aligning to protein-coding gene bodies were extracted, and read counts were normalized. cfDNA tumor fraction was estimated using IchorCNA. Machine learning models were trained using k-fold cross-validation and confounder-based cross-validations to assess generalization performance. RESULTS: In a colorectal cancer cohort heavily weighted towards early-stage cancer (80% stage I/II), we achieved a mean AUC of 0.92 (95% CI 0.91-0.93) with a mean sensitivity of 85% (95% CI 83-86%) at 85% specificity. Sensitivity generally increased with tumor stage and increasing tumor fraction. Stratification by age, sequencing batch, and institution demonstrated the impact of these confounders and provided a more accurate assessment of generalization performance. CONCLUSIONS: A machine learning approach using cfDNA achieved high sensitivity and specificity in a large, predominantly early-stage, colorectal cancer cohort. The possibility of systematic technical and institution-specific biases warrants similar confounder analyses in other studies. Prospective validation of this machine learning method and evaluation of a multi-analyte approach are underway.

Machine learning for classifying tuberculosis drug-resistance from DNA sequencing data.

Motivation: Correct and rapid determination of Mycobacterium tuberculosis (MTB) resistance against available tuberculosis (TB) drugs is essential for the control and management of TB. Conventional molecular diagnostic test assumes that the presence of any well-studied single nucleotide polymorphisms is sufficient to cause resistance, which yields low sensitivity for resistance classification. Summary: Given the availability of DNA sequencing data from MTB, we developed machine learning models for a cohort of 1839 UK bacterial isolates to classify MTB resistance against eight anti-TB drugs (isoniazid, rifampicin, ethambutol, pyrazinamide, ciprofloxacin, moxifloxacin, ofloxacin, streptomycin) and to classify multi-drug resistance. Results: Compared to previous rules-based approach, the sensitivities from the best-performing models increased by 2-4% for isoniazid, rifampicin and ethambutol to 97% (P < 0.01), respectively; for ciprofloxacin and multi-drug resistant TB, they increased to 96%. For moxifloxacin and ofloxacin, sensitivities increased by 12 and 15% from 83 and 81% based on existing known resistance alleles to 95% and 96% (P < 0.01), respectively. Particularly, our models improved sensitivities compared to the previous rules-based approach by 15 and 24% to 84 and 87% for pyrazinamide and streptomycin (P < 0.01), respectively. The best-performing models increase the area-under-the-ROC curve by 10% for pyrazinamide and streptomycin (P < 0.01), and 4-8% for other drugs (P < 0.01). Availability and implementation: The details of source code are provided at http://www.robots.ox.ac.uk/~davidc/code.php. Contact: david.clifton@eng.ox.ac.uk. Supplementary information: Supplementary data are available at Bioinformatics online.

Machine Learning and Decision Support in Critical Care.

Clinical data management systems typically provide caregiver teams with useful information, derived from large, sometimes highly heterogeneous, data sources that are often changing dynamically. Over the last decade there has been a significant surge in interest in using these data sources, from simply re-using the standard clinical databases for event prediction or decision support, to including dynamic and patient-specific information into clinical monitoring and prediction problems. However, in most cases, commercial clinical databases have been designed to document clinical activity for reporting, liability and billing reasons, rather than for developing new algorithms. With increasing excitement surrounding "secondary use of medical records" and "Big Data" analytics, it is important to understand the limitations of current databases and what needs to change in order to enter an era of "precision medicine." This review article covers many of the issues involved in the collection and preprocessing of critical care data. The three challenges in critical care are considered: compartmentalization, corruption, and complexity. A range of applications addressing these issues are covered, including the modernization of static acuity scoring; on-line patient tracking; personalized prediction and risk assessment; artifact detection; state estimation; and incorporation of multimodal data sources such as genomic and free text data.

Whole-genome sequencing for prediction of Mycobacterium tuberculosis drug susceptibility and resistance: a retrospective cohort study.

BACKGROUND: Diagnosing drug-resistance remains an obstacle to the elimination of tuberculosis. Phenotypic drug-susceptibility testing is slow and expensive, and commercial genotypic assays screen only common resistance-determining mutations. We used whole-genome sequencing to characterise common and rare mutations predicting drug resistance, or consistency with susceptibility, for all first-line and second-line drugs for tuberculosis. METHODS: Between Sept 1, 2010, and Dec 1, 2013, we sequenced a training set of 2099 Mycobacterium tuberculosis genomes. For 23 candidate genes identified from the drug-resistance scientific literature, we algorithmically characterised genetic mutations as not conferring resistance (benign), resistance determinants, or uncharacterised. We then assessed the ability of these characterisations to predict phenotypic drug-susceptibility testing for an independent validation set of 1552 genomes. We sought mutations under similar selection pressure to those characterised as resistance determinants outside candidate genes to account for residual phenotypic resistance. FINDINGS: We characterised 120 training-set mutations as resistance determining, and 772 as benign. With these mutations, we could predict 89·2% of the validation-set phenotypes with a mean 92·3% sensitivity (95% CI 90·7-93·7) and 98·4% specificity (98·1-98·7). 10·8% of validation-set phenotypes could not be predicted because uncharacterised mutations were present. With an in-silico comparison, characterised resistance determinants had higher sensitivity than the mutations from three line-probe assays (85·1% vs 81·6%). No additional resistance determinants were identified among mutations under selection pressure in non-candidate genes. INTERPRETATION: A broad catalogue of genetic mutations enable data from whole-genome sequencing to be used clinically to predict drug resistance, drug susceptibility, or to identify drug phenotypes that cannot yet be genetically predicted. This approach could be integrated into routine diagnostic workflows, phasing out phenotypic drug-susceptibility testing while reporting drug resistance early. FUNDING: Wellcome Trust, National Institute of Health Research, Medical Research Council, and the European Union.

Phenotypic characterisation of Crohn's disease severity.

Crohn's disease (CD) is a highly heterogeneous disease, with great variation in patient severity. Using supervised machine learning techniques to predict severity from common laboratory and clinical measurements, we found that high levels of C-reactive protein and low levels of lymphocytes and albumin are important predictive factors. Building upon this knowledge, we used extreme value theory to create a probabilistic model that combines information about behaviour in the extremes of these lab measurements to produce a single risk score over time. We then clustered these patient risk scores to identify several common clinical trajectories for CD patients.

First steps in using machine learning on fMRI data to predict intrusive memories of traumatic film footage.

After psychological trauma, why do some only some parts of the traumatic event return as intrusive memories while others do not? Intrusive memories are key to cognitive behavioural treatment for post-traumatic stress disorder, and an aetiological understanding is warranted. We present here analyses using multivariate pattern analysis (MVPA) and a machine learning classifier to investigate whether peri-traumatic brain activation was able to predict later intrusive memories (i.e. before they had happened). To provide a methodological basis for understanding the context of the current results, we first show how functional magnetic resonance imaging (fMRI) during an experimental analogue of trauma (a trauma film) via a prospective event-related design was able to capture an individual's later intrusive memories. Results showed widespread increases in brain activation at encoding when viewing a scene in the scanner that would later return as an intrusive memory in the real world. These fMRI results were replicated in a second study. While traditional mass univariate regression analysis highlighted an association between brain processing and symptomatology, this is not the same as prediction. Using MVPA and a machine learning classifier, it was possible to predict later intrusive memories across participants with 68% accuracy, and within a participant with 97% accuracy; i.e. the classifier could identify out of multiple scenes those that would later return as an intrusive memory. We also report here brain networks key in intrusive memory prediction. MVPA opens the possibility of decoding brain activity to reconstruct idiosyncratic cognitive events with relevance to understanding and predicting mental health symptoms.

Risk of cardiovascular disease from antiretroviral therapy for HIV: a systematic review.

BACKGROUND: Recent studies suggest certain antiretroviral therapy (ART) drugs are associated with increases in cardiovascular disease. PURPOSE: We performed a systematic review and meta-analysis to summarize the available evidence, with the goal of elucidating whether specific ART drugs are associated with an increased risk of myocardial infarction (MI). DATA SOURCES: We searched Medline, Web of Science, the Cochrane Library, and abstract archives from the Conference on Retroviruses and Opportunistic Infections and International AIDS Society up to June 2011 to identify published articles and abstracts. STUDY SELECTION: Eligible studies were comparative and included MI, strokes, or other cardiovascular events as outcomes. DATA EXTRACTION: Eligibility screening, data extraction, and quality assessment were performed independently by two investigators. DATA SYNTHESIS: Random effects methods and Fisher's combined probability test were used to summarize evidence. FINDINGS: Twenty-seven studies met inclusion criteria, with 8 contributing to a formal meta-analysis. Findings based on two observational studies indicated an increase in risk of MI for patients recently exposed (usually defined as within last 6 months) to abacavir (RR 1.92, 95% CI 1.51-2.42) and protease inhibitors (PI) (RR 2.13, 95% CI 1.06-4.28). Our analysis also suggested an increased risk associated with each additional year of exposure to indinavir (RR 1.11, 95% CI 1.05-1.17) and lopinavir (RR 1.22, 95% CI 1.01-1.47). Our findings of increased cardiovascular risk from abacavir and PIs were in contrast to four published meta-analyses based on secondary analyses of randomized controlled trials, which found no increased risk from cardiovascular disease. CONCLUSION: Although observational studies implicated specific drugs, the evidence is mixed. Further, meta-analyses of randomized trials did not find increased risk from abacavir and PIs. Our findings that implicate specific ARTs in the observational setting provide sufficient evidence to warrant further investigation of this relationship in studies designed for that purpose.